2016 26th International Telecommunication Networks and Applications Conference (ITNAC)
Securing RPL Routing Protocol from Blackhole
Attacks Using a Trust-based Mechanism
David Airehrour, Jairo Gutierrez
Auckland University of Technology
Auckland, New Zealand
{dairehrour, jairo.gutierrez}@aut.ac.nz
Abstract—This research addresses blackhole routing attacks
—a fundamental security attack on the routing of data in IoT
networks. Most IoT devices today, from medical devices to
connected vehicles and even smart buildings, come bundled with
the capability of communicating wirelessly with one another.
Consumers are progressively embracing the concept of connected
devices while recent studies indicate that security is not high on
the priority list of manufacturers especially in the way these
devices route and communicate data amongst themselves thus,
leaving the door wide open to attacks and compromises. We
propose a trust-based RPL routing protocol, which addresses the
blackhole attacks. We show that our proposed system is both
secure from blackhole attacks while not imposing undue
overheads on network traffic
Keywords— IoT; IIoT; DODAG; RPL; DIO; DAO; DIS;
MRHOF; Trust; Blackhole; Contiki/Cooja
I. INTRODUCTION
The Internet of Things (IoT) has been recognized as a trend
causing a global technological disruption today and as a result
of a melding of advances in computing and communication
enterprises. IoT is set to transform not only the user-to-machine
interaction but also the manner in which machine-to-machine
interacts. Already, we are witnessing the penetration of IoT
devices in the market place. Various industrial sectors have
begun witnessing the permeation of IoT products into the
fabric of several industries including healthcare, energy,
automotive and agriculture. Increasingly in these industries,
users are witnessing the Industrial Internet of Things (IIoT),
where devices such as sensors, exercise fit bits, robots, and
insulin pumps are progressively becoming more connected to
one another. It is perceived that Internet of Things will
significantly change the future of not only the industrial sectors
of the world but also will bring a positive transformation to
how we live. A culmination of the full potential of the IoT
vision will better the standards of living of humanity because
of the numerous value creation opportunities while also
improving the careers of many. It is expected that the adoption
of IoT will culminate in a web of new smart paradigms like
smart healthcare, smart agriculture and smart power. This
could eventually evolve into new ecosystems of IoT that are
propelled by self-aware, autonomic machines.
However, the fact that these devices can communicate with
one another and over the web, poses a security risk to even the
Industrial Control Systems (ICSs) and other connected online
978-1-5090-0919-0/16/$31.00 ©2016 IEEE
Sayan Kumar Ray
Manukau Institute of Technology
Auckland, New Zealand
sayan.ray@manukau.ac.nz
devices and hence, this requires better security mechanisms.
There is no doubt that IoT is creating a new epoch of
innovation that connects the digital and machine ecosystems
and brings better speed and effectiveness to many sectors as
recounted above. Nevertheless, with sensitive information
increasingly being made available online via the deployment of
IoT, and more endpoints exposed to attackers, the research
community and indeed the business world are swiftly
recognizing that security cannot be an afterthought.
A study by McKinsey [1] projects that the cost of
cybersecurity will increase to $3 trillion by 2020 and of which,
many are futile. Further to the projection by Ericsson [2] that
the number of connected devices will reach 50 billion by 2020,
there is a pressing need to profoundly rethink security for the
always connected, high-volume and distributed world of the
Internet of Things. One typical area of exposure in IoT is the
transmission of routing packets among IoT devices. These
packets move across heterogeneous networks and are thus
susceptible to various attacks common to both the digital and
machine world. At this stage of the nascent development of
IoT, the security challenges need to be addressed in order to
engender confidence in the public and globally achieve success
with IoT.
The objective of this research is to develop a lightweight
trust-based Routing Protocol for low power and Lossy
networks (RPL) that will address blackhole attacks in IoT. The
rest of the paper is organized as follows. Section II discusses
the IoT routing protocols and the current industry standards.
Section III introduces the security features available in RPL
and highlights the challenges in its implementation. A trust-
based mechanism for RPL routing protocol is introduced in
Section IV. We show that our proposed system is both secure
from blackhole attacks while also not imposing undue
overheads on network traffic. In section V, we present our
simulation results using the Contiki/Cooja environment and we
demonstrate the efficacy of our proposed trust-based RPL
routing protocol. Finally, the conclusions and final notes on our
future work are articulated in Section VI.
II. INTERNET OF THINGS: A ROUTING PROTOCOL
PERSPECTIVE
A. Routing Protocols in IoT
A routing protocol is a communication system tasked with
the responsibility of making intelligent routing decisions during
 2016 26th International Telecommunication Networks and Applications Conference (ITNAC)
the forwarding of routing data among nodes. In sensor
networks, there are two types of routing namely: proactive and
reactive routing and protocols developed are based on any of
these two systems.
B. Routing Protocol for Low-power and Lossy Networks
The Routing protocol for low power and lossy networks
(RPL) is an IPv6 routing protocol designed by the IETF group
known as Routing Over Low power and Lossy networks
(ROLL) [3]. The protocol was designed as a standard for low
power and lossy networks, which includes all IoT sensor
nodes. RPL is a proactive routing protocol, which operates by
discovering routes as soon as the RPL network is started. It
forms a tree-like topology known as Destination Oriented
Directed Acyclic Graph (DODAG). Every node in the RPL
network selects a preferred parent based on some metrics and
this preferred parent acts like a gateway for that node. If a node
seeks to forward a packet for which it does not have a path in
its routing table, it simply forwards it to its preferred parent
who has a path either to the destination or to its own parent for
onward transmission until it gets to the final destination in the
tree. Path selection is an important factor for RPL and hence
uses multiple metrics for this purpose. Every node in the
DODAG computes its rank from the perspective of the position
of the DODAG root node (sink) and in relation to the position
of the other nodes. The rank of a node decreases in the upward
direction towards the DODAG root while it increases from the
DODAG root towards the leaf nodes (sender nodes). RPL
operates in two modes in order to perform downward routing
and they are RPL non-storing mode (source routing) and RPL
storing mode (stateful in-network routing). In storing mode,
each packet holds the route path to the destination. This entails
the DODAG root maintaining details about each node within
the network. It is important to note that when operating in a
non-storing mode, forwarding RPL nodes in the network need
to retain their in-network routing tables in order to identify
where to send their packets. However, in both modes discussed
above, the RPL DODAG root still retains a database of all
nodes for downward routing purpose [3].
RPL utilizes three control message types for the creation
and maintenance of its graph topology and route table. The
control messages include: DODAG Information Object (DIO),
DODAG Advertisement Object (DAO) and DODAG
Information Solicitations (DIS). DIO is used for creation,
maintenance and discovery of the DODAG topology. When a
RPL network is started nodes exchange DODAG information
via the DIO. The DIO helps nodes to select their preferred
parents. RPL uses DAO messages to transmit the prefix of a
node to its ancestor nodes for downward routing purposes. The
DIS message is used by any unattached node in the network to
solicit for a potential parent node. DIS is triggered by a node in
a situation when it cannot obtain a DIO after a certain time
interval [3].
III. SECURITY IN RPL
Security has been identified as being critical in low power
and lossy networks [4]. In addition, the complexity of
deployment and size is also a core concern for these resource-
constrained networks such that it may be economically, if not
physically impossible, to embed sophisticated security
mechanisms in a RPL implementation. Further to that, many
deployments can utilize link-layer or other security
mechanisms to meet their security requirements without
requiring the use of security in RPL. Consequently, the security
features in RPL come in as optional extensions. When in
operation, RPL nodes could function in three security modes.
In the first, referred to as the "unsecured", RPL control
messages are forwarded without any extra security primitives.
The unsecured mode infers that the RPL network may well be
using other security mechanisms (such as a link-layer security)
in order to meet application specific requirements. In the
second, referred to as "pre-installed", nodes entering a RPL
instance have pre-installed keys that permit them to process,
generate and safeguard RPL messages. In the third mode,
referred to as "authenticated", nodes can enter as leaf nodes by
means of pre-installed keys in pre-installed mode, or enter as
forwarding nodes by getting a key from a central authentication
authority. In the last two modes, every RPL message has a
secure variant. The security level (32-bit and 64-bit MAC and
ENC-MAC modes are all supported) and the algorithms (CCM
and AES-128 are also supported) in operation are specified in
the protocol messages. The secure variants of the RPL
messages are meant to provide confidentiality, integrity, delay
protection and replay protection as an added option. However,
the bad news is they all rely on past encryption solutions that
have failed —and which continue to fail [5]. Public Key
Infrastructure (PKI) was developed about four decades ago to
safeguard the communications between two human parties. It
was at no time designed to handle the complications of
managing industrial-scale networks of 50 billion devices that
IoT promises to usher in. The very thought of having a central
authentication authority for billions of devices makes it
extremely awkward and inefficient.
A. Attacks in RPL
The RPL protocol, like any other wireless sensor network
protocol, has been shown to be vulnerable to routing attacks.
These attacks have been researched and covered in [6-8]
among other papers; Table 1 shows a summary of attacks in
RPL and some proposed solutions.
We examine with special interest the blackhole attack
solutions proposed in table 1. Reference [9] assumes the use of
cryptography and it specifically uses Secure Hash Algorithm 1
(SHA-1) as the hash function to protect the route messages
being transmitted. It also assumes that the cryptographic
system utilized is guaranteed hence, it will not be tampered
with by any malicious nodes. As established in section III, the
use of cryptography (SHA-1) will certainly deplete the battery
energy of the nodes and hence degrade network performance.
The assumption that the attacking nodes will not tamper
with the cryptographic system makes the proposed solution
impracticable in a real world scenario. Of equal importance is
the mobility of the nodes, when these nodes join and leave the
network at will, implementing encryption becomes difficult as
a specific node with certain network details required by other
nodes suddenly becomes unavailable. The authors of [10]
revealed the weaknesses in the implementation of the
ContikiRPL viz-a-viz malicious attacks
 2016 26th International Telecommunication Networks and Applications Conference (ITNAC)

TABLE I. SUMMARY OF RPL ATTACKS AND COUNTERMEASURES

Type of attack Consequence on performance of network Some proposed solutions
Rank Minimal packet delivery and high packet loss; non- Use of IDS based solutions [11,12], VeRA [13], TRAIL[14]
optimal path selection and routing loop
Selective forwarding Destabilization of route topology Heartbeat protocol [6]
Sinkhole Transmitting network traffic via attacker node IDS solution [11], Parent fail-over, rank authentication technique
[9]
Hello flooding Degrading of sensor energy RPL’s global and local repair mechanism removes attack
Wormhole Destabilization of route topology and network traffic Merkle tree authentication [15]
Sybil and Route traffic unreachable to node destination Routing attacks and countermeasures in RPL-Based IoT [6]
Clone ID
Denial of Service Unavailability of network resources Intended user IDS based solution [16]
Blackhole High packet drop rate and high control and route SVELTE [11]
traffic overhead Monitoring of counters [8], Parent fail-over [9],
Version number High traffic latency and high control overhead with VeRA [13]
minimal packet delivery ratio;
Local repair Route and control traffic destabilization. IDS based solution [4]
Control overhead
Neighbour and DIS Falsification of route and network resource depletion, TRAIL [14]
attacks Network resource depletion

and thus, gave helpful insight into design issues that could help
in the implementation of a better ContikiRPL. The authors of
[11] implemented an IDS system to defend against sinkhole
and selective forwarding attacks and opined that it could also
detect blackhole attacks; however, they assumed that key IDS
nodes must be strategically placed. With a deluge of IoT
devices randomly and remotely located, this may not be the
case and thus, may not provide optimal defense against attacks.
We present in section IV an algorithmic trust-based
approach to secure the RPL routing protocol. This proposed
system, when implemented in RPL, counters blackhole attacks.
IV. A TRUST-BASED MECHANISM FOR RPL
Blackhole attacks perform malicious activities like causing
high packet drops and high route and control packet overhead,
which depletes the limited resources of the IoT nodes. When
malicious nodes propagate blackhole attacks, network latency
increases and the ranks of the nodes are altered, which causes a
disruption to the RPL network topology and to its stability.
Additionally, the rank alteration causes the nodes to re-
compute their ranks. The rank alteration triggers a local repair
—a self-healing mechanism that RPL uses to eliminate local
routing loops. However, with the increase in blackhole attacks,
the local repair eventually becomes inefficient prompting a
global repair by the DODAG root. A continuous initiation of
these repair messages causes inefficiencies and disruption to
the RPL network.
As argued in section III, the security related solutions to
prevent malicious activities in RPL, which include
cryptography and authentication operations, are unable to cope
with the billions of IoT devices. Besides, the encryption
technology could be considered complex and energy
consuming in the context of the limited available resources of
the IoT sensor nodes. Therefore, a trust-based mechanism
which employs a lightweight solution with respect to the
limited resources of the nodes presents an interesting solution
for the security of RPL routing.
A. Embedding Trust in RPL
We describe below our proposed trust-based mechanism.
The aim of this mechanism is to compute a trust value for each
node in the RPL network while embedding computed trust
values for routing decisions. This way it will deliver the
combined values of providing an optimal routing decision
while also isolating malicious nodes that may seek to drop
control and route packets. The trust mechanism also computes
the effective feedbacks values between nodes. In our model,
we make two basic assumptions. i) That every node operates
in promiscuous mode hence, they can overhear neighbour
packet transmissions. ii) That every blackhole attacking node
will overtime begin to drop all route packets thus, the effective
feedbacks between nodes (i.e. the number of packets a node
was able to satisfactorily forward on behalf of the requesting
node) will certainly reflect the blackhole nature of any node.
In our new protocol, a trust-based mechanism is embedded
into RPL to enhance its capability to isolate blackhole attacks.
When RPL is initially started, a comparison is made
between nodes based on the expected transmission count and
the rank of the nodes. These are normal RPL operations to
determine preferred parents and routing decisions. Further to
that, our computed trust values as depicted in equation 1 are
sorted in descending order of magnitude of trust. The
corresponding trusted node(s) are selected for routing
decisions while still maintaining the rank order of all nodes in
the RPL network. The trust is computed as:
(1)
Where Ndlv is the number of node i’s packets delivered through
node j and Nsent is the total number of packets sent by node i to
node j. Our trust-based algorithm is shown in Fig. 1.
RPL uses routing metrics defined in its Objective Function
to create the DODAG. Essentially, the routing metrics defined
in the objective function helps in the creation of the network
 2016 26th International Telecommunication Networks and Applications Conference (ITNAC)

Algorithm for blackhole detection TABLE II. TABLE STYLES

Let N1 ← one unfilled node in the NeighborNodeList [ ]
Let N2 ← another node next to N1 in the NeighborNodeList [ ] Simulation Parameters
Simulation tool Contiki/Cooja 3.0
Compute Mote type Tmote Sky
Simulation run time 3600 seconds
If (N1.ETX<= ETX_Limit) & (N2.ETX<=ETX_Limit) Simulation coverage area 70m x 70m
If (N1.Rank <= Rank_Self) & (N2.RANK <= RANK_Self) Interference range 100m
Preferred_Parent = N1.EP > N2.EP ? N1 : N2;
Else Total number of nodes 30
If (N1.Rank <= Self_Rank) || (N2.Rank <= Self_Rank) Root node (sink) 1
Preferred_Parent = N1.Rank < N2.Rank ? N1 : N2 Blackhole attack nodes 3
Else
Preferred_Parent = NULL; Legitimate nodes 26
Else Deployment environment Smart building
If (N1.ETX <= ETX_Limit) || (N2.ETX <= ETX_Limit) Wireless transmission range 50 metres
Preferred_Parent = N1.ETX <= N2.ETX ? N1 : N2;
Else Network protocol IP based
Preferred_Parent = NULL; Routing protocol RPL
Return Preferred_Parent
End program.
Fig. 1. A trust-based algorithm for the isolation of malicious nodes in RPL

routes and hence, resulting in an optimal route. In the Contiki

implementation of RPL, there are 2 objective functions
namely: Minimum Rank with Hysteresis Objective Function
(MRHOF) based on RFC 6719 [17] and Objective Function
zero (OF0). Contiki uses MRHOF by default, which minimizes
the expected transmission count (ETX) values. This research
work compares the MRHOF’s implementation of RPL with our
trust-based implementation of RPL.

V. SIMULATION AND RESULTS

In the simulation, we have assumed that the IoT sensors are
deployed in a smart building with one level. The InstantContiki
3.0 platform [18] is used to perform the simulation while
embedding the trust-based mechanism (shown in section IV)
into RPL (MRHOF). The various simulation parameters are Fig. 2. A network topology view of the IoT sensor nodes
listed in table II. During simulation, the system takes into
account interference from its surrounding such as other devices
or technologies that may be in use. We have also used the
TMote Sky mote (Cooja simulator) for simulation with
transmission range set to 50 metres and the interference range
to 100 metres.
A topology arrangement of the sensor nodes is displayed in
Fig. 2 while the blackhole attack was manually activated
during RPL operation. In the simulation study, we have Fig. 3. A sequence of packets sent and received by the sender and sink nodes
assumed that the attack nodes behave as good nodes from
inception and commence their malicious activities over time eliminated lossy network links because we made nodes to have
(when activated). strong reachability to their neighbours. Furthermore, we have
examined the packets dropped by the malicious nodes and they
A. Isolation of Blackhole Attacks corresponded to the packets that have failed to reach the sink
In the simulation, sender nodes transmit packets to the sink node. A complete log of the sent and received packets was
node with the following stamp on each packet sent: time, analyzed and the results presented in Fig. 5. In Fig. 4, the trust-
source ID, packet type (sent or received), destination ID, based RPL protocol was able to detect and isolate the blackhole
sequence number and data size. This is shown in Fig 3. Packet attacks during routing operations. A highlight of the attacks
sequence IDs are matched to ensure that packets sent are detected can be seen from the encircling blue pen-mark. In
received by the sink node. Any sent packet sequence ID that is addition, Fig. 5 displays a graph summary of attacks detected
not matched with a corresponding received sequence ID by the and isolated during RPL operation using the trust-based RPL
sink node has either been black holed by the malicious node or mechanism over a 60-minute simulation period at an interval of
affected by the lossy network link. However, we have every 5 minutes. As many as 600 attacks were detected
between the 40th and 45th minute of the RPL operation.
 2016 26th International Telecommunication Networks and Applications Conference (ITNAC)

Conversely, in MRHOF's RPL implementation these attacks

could not be detected, as there was no mechanism to detect nor
isolate blackhole attacks.
It is of note that in RPL routing, a node rank change shows
a re-alignment of a child-node to another preferred parent-
node. Blackhole attack nodes advertise themselves to their
neighbour nodes as better routes in a guise to attract these
unsuspecting nodes while eventually dropping their packets. In
Fig. 6, a comparison of the frequency of node rank changes
between the two routing systems is made. RPL with MRHOF
showed high frequency in rank changes reflecting its high level
of susceptibility to blackhole attacks while our trust-based RPL
protocol showed a very marginal level of susceptibility.

B. Network Performance
Even though we have a system in place which could detect
and isolate blackhole attacks during RPL operations, it
becomes imperative that the new system should not impose
undue overhead on the network performance. We present
below a measurement of network throughput and packet loss Fig. 4. Detection of Blackhole attacking nodes during RPL operation
rate to determine if our proposed system is able to deliver
reasonable levels of network performance while isolating
blackhole attacks when compared to MRHOF’s RPL.
In Fig. 7, the trust-based RPL showed significant
improvement in throughput over the standard RPL (MRHOF).
In fact, the throughput measurement of nodes 2-9, 15, 18, 19,
20, 22 and 25 was 0 kbps under MRHOF’s RPL as a result of
the blackhole attacks on the network. This indicates that these
nodes were child-nodes to a blackhole parent-node.
Meanwhile, with the trust-based RPL protocol, none of the
nodes had a throughput of 0 kbps which implies that no child
node had a blackhole parent node.
Fig. 8 displays a graphical representation of the percentage
of packet loss in RPL routing operation under blackhole
attacks. While the trust-based RPL protocol's packet loss Fig. 5. Trust-based detection and isolation of blackhole attacks in RPL
stayed below 40%, the standard RPL (MRHOF) recorded a
staggering 60 to 100% packet loss rate. Thus, the two network
performance measurements presented above justify the trust-
based RPL routing protocol as a better performing protocol
over the standard RPL (MRHOF) under blackhole attacks.

VI. CONCLUSIONS
In IoT networks, compromised sensor nodes can destabilize
the integrity of routing data by (a) intentionally sending
incorrect control and route information, (b) dropping all
packets, (c) injecting false routing data during data
aggregation, and (d) hampering the forwarding of composited
data. Since cryptographic methods have proved to be
inadequate in the prevention of these attacks, especially on a
massive scale of billions of IoT nodes, a trust-based Fig. 6. Comparison of frequency of node rank changes during blackhole
mechanism has been proposed. This research has proposed a attacks in RPL network during simulation
new reliable routing protocol that provides a feedback-aware
trust-based security system for IoT networks. The proposed research work provides a comprehensive security solution
system computes a trust value for any node based on the good against blackhole attacks.
forwarding behavior of neighbour nodes in the network. The
trust value is dependent on the positive feedback awareness Our future work intends to incorporate energy metrics into
among the nodes and the trust evaluation analysis. As the system, so that it could observe the nodes with depleted
demonstrated in the obtained simulation results, the proposed
 2016 26th International Telecommunication Networks and Applications Conference (ITNAC)
energy levels and isolate them from routing decisions, while
providing them with the opportunity to recoup their battery
power.
Fig. 7. Comparison of throughput measurements between RPL (MRHOF)
and Trust-based RPL
Fig. 8. Packet loss rate comparison between RPL (MRHOF) and Trust-based
RPL
REFERENCES
[1] D. Chinn, J. Kaplan, and A. Weinberg, "Risk and responsibility in a
hyperconnected world: Implications for enterprises," McKinsey Global
InstituteJanuary 2014 2014.
[2] Ericsson. (2011, December 2, 2014). More than 50 billion connected
devices: Driving forces. Available: http://www.akos-
rs.si/files/Telekomunikacije/Digitalna_agenda/Internetni_protokol_Ipv6/
More-than-50-billion-connected-devices.pdf
[3] T. Winter, P. Thubert, A. Brandt, J. Hui, R. Kelsey, P. Levis, et al.
(2012, RPL: IPv6 Routing Protocol for Low-Power and Lossy
Networks. Available: https://tools.ietf.org/html/rfc6550
[4] A. Le, J. Loo, A. Lasebae, M. Aiash, and Y. Luo, "6LoWPAN: a study
on QoS security threats and countermeasures using intrusion detection
system approach," International Journal of Communication Systems,
vol. 25, pp. 1189-1212, 2012.
[5] A. Nordrum. (2016) Quantum Computer Comes Closer to Cracking
RSA Encryption. IEEE Spectrum. Available:
http://spectrum.ieee.org/tech-
talk/computing/hardware/encryptionbusting-quantum-computer-
practices-factoring-in-scalable-fiveatom-experiment
[6] L. Wallgren, S. Raza, and T. Voigt, "Routing Attacks and
Countermeasures in the RPL-Based Internet of Things," International
Journal of Distributed Sensor Networks, vol. 2013, p. 11, 2013.
[7] T. Tsao, R. Alexander, M. Dohler, V. Daza, A. Lozano, and M.
Richardson, "A Security Threat Analysis for Routing Protocol for Low-
power and lossy networks (RPL)," 2014.
[8] K. Chugh, L. Aboubaker, and J. Loo, "Case Study of a Black Hole
Attack on LoWPAN-RPL," in Proc. of the Sixth International
Conference on Emerging Security Information, Systems and
Technologies (SECURWARE), Rome, Italy (August 2012), 2012, pp.
157-162.
[9] K. Weekly and K. Pister, "Evaluating sinkhole defense techniques in
RPL networks," in Network Protocols (ICNP), 2012 20th IEEE
International Conference on, 2012, pp. 1-6.
[10] K. Chugh, A. Lasebae, and J. Loo, "Case Study of a Black Hole Attack
on 6LoWPAN-RPL," in SECURWARE 2012: The Sixth International
Conference on Emerging Security Information, Systems and
Technologies Rome, Italy, 2012.
[11] S. Raza, L. Wallgren, and T. Voigt, "SVELTE: Real-time intrusion
detection in the Internet of Things," Ad Hoc Netw., vol. 11, pp. 2661-
2674, 2013.
[12] S. O. Amin, M. S. Siddiqui, C. S. Hong, and J. Choe, "A novel coding
scheme to implement signature based IDS in IP based Sensor
Networks," in Integrated Network Management-Workshops, 2009.
IM'09. IFIP/IEEE International Symposium on, 2009, pp. 269-274.
[13] A. Dvir, T. Holczer, and L. Buttyan, "VeRA-version number and rank
authentication in rpl," in Mobile Adhoc and Sensor Systems (MASS),
2011 IEEE 8th International Conference on, 2011, pp. 709-714.
[14] H. Perrey, M. Landsmann, O. Ugus, T. C. Schmidt, and M. Wählisch,
"TRAIL: Topology Authentication in RPL," arXiv preprint
arXiv:1312.0984, 2013.
[15] K. Zhang, X. Liang, R. Lu, and X. Shen, "Sybil Attacks and Their
Defenses in the Internet of Things," Internet of Things Journal, IEEE,
vol. 1, pp. 372-383, 2014.
[16] P. Kasinathan, C. Pastrone, M. Spirito, and M. Vinkovits, "Denial-of-
Service detection in 6LoWPAN based Internet of Things," in Wireless
and Mobile Computing, Networking and Communications (WiMob),
2013 IEEE 9th International Conference on, 2013, pp. 600-607.
[17] O. Gnawali, "The minimum rank with hysteresis objective function,"
2012.
[18] Thingsquare. (2016, June, 2016). Contiki: The Open Source OS for the
Internet of Things, Available: http://www.contiki-os.org/download.html